AI has made it easier than ever for anyone to create content. Blog posts, landing pages, emails, social posts, etc. What used to take hours can now take minutes. And we’re seeing the result of this with everyone posting a whole lot of content, all the time.
But there’s a difference between producing more content and producing better content (and even producing great content that provides value). As AI providers and LLMs introduce new ways to identify AI-generated text, that distinction is becoming even more important.
Anthropic, the company behind Claude, announced invisible text watermarking for supported models in August 2026. OpenAI announced its own text watermarking approach in October.
Neither approach adds a visible watermark to the text. Instead, the watermark creates a machine-detectable pattern in the model’s word choices. It can remain detectable after copying and pasting, although editing and other limitations can affect detection.
This has understandably raised a lot of questions. Can search engines identify AI-generated content? Will using AI hurt your rankings? And if AI can write just about anything, how do you make sure your content stands out?
The issue isn’t even the watermark. It’s that all content looks and sounds the same.
So how do you create content that stands out and earns attention in an increasingly AI-powered world?
TL;DR:
- Watermarking helps identify whether a particular AI system may have generated text. It doesn’t automatically make that content bad or determine whether it deserves to rank.
- AI detection and watermarking aren’t new concepts. Different LLMs use different approaches, and not every LLM watermarking system works the same way.
- Google’s concern is low-value content, not AI itself. Producing large volumes of content and pages without adding value can violate spam policies.
- Originality is your competitive advantage. Your expertise, customer insights, proprietary data, and point of view are what make content valuable to your audience.
- Make content quotable, not just crawlable. Give people and AI-powered search experiences something specific, credible, and useful to reference.
What is AI content watermarking?
An AI watermark is a signal embedded in content that can help a detection system determine whether a particular AI model likely generated the content. Unlike a traditional watermark on an image, a text watermark doesn’t need to be visible to the person reading the content.
For example, Anthropic subtly influences certain word choices during text generation, creating a pattern that a compatible detector can see. OpenAI has announced a different approach, called textGrain, that also uses subtle patterns in word choices. These approaches are part of a broader movement toward understanding where digital content came from and how it was created.
This isn’t limited to just Claude and ChatGPT. AI developers and the broader technology industry have been exploring different ways to identify AI-generated content, including watermarks, metadata, and more. However, it’s important to understand the distinction between those technologies. Not every LLM embeds a detectable text watermark, and a detector designed for one provider’s watermark may not recognize another’s.
A watermark also doesn’t automatically tell you whether content is accurate, useful, original, or written entirely by AI.
Will AI watermarks hurt your SEO rankings?
Google has said that appropriate use of AI to create content isn’t against its guidelines. Its focus is on the quality and purpose of the content. However, they do warn against scaled content abuse, which includes producing large amounts of content primarily to manipulate search rankings rather than help users. That policy applies whether the content is generated by AI, humans, or both.
In other words, a watermark isn’t going to cause you to rank lower in Google search results.
There is nothing that proves simply having an AI watermark causes a page to lose rankings, but that doesn’t mean every AI-generated page is safe from ranking problems. Content that offers little originality, little added value, or an unsatisfactory experience can still struggle.
Consider two companies publishing content about the same topic:
Company A uses AI to generate 20 blog posts, makes minor edits, and publishes them all in a day.
Company B uses AI to organize research and develop outlines. Its team adds customer insights, original analysis, relevant examples, expert perspectives, and clear recommendations.
Both companies use AI. The difference is what they contribute.
The second company has a much stronger foundation for creating useful content because it brings something beyond what a model can generate from a generic prompt.
That is what matters here. As people increasingly use AI and LLMs to research products, compare vendors, and answer business questions, your content needs to offer useful information worth referencing. Don’t waste time just trying to hide a watermark or make your content appear less AI-generated. Spend that time making the content more valuable.
The real problem with using AI for content
This is where I think we all need to pay closer attention. AI can produce a perfectly acceptable blog post about almost any topic. It can summarize best practices, explain common challenges, and assemble a list of recommendations. But if everyone is using similar tools to answer the same questions with the same information, what makes one company’s content more valuable than another’s?
Not much.
Let’s say you’re a B2B marketing agency writing about email marketing benchmarks for 2027. You could prompt AI to write an article titled, “Email Marketing Benchmarks for 2027.” It might summarize industry averages, explain open rates and click-through rates, and offer some tips for improving performance. There’s nothing inherently wrong with that. But how many dozens of other companies could create something similar?
Now, imagine you actually analyze 14,000 emails and publish an article titled, “We Analyzed 14,000 Emails. Here’s What We Learned.” Suddenly, you have something different. You have specific data. You have original findings. You can explain what you observed, what surprised you, what the numbers mean, and what other marketers should do with that information.
You could use AI to help you analyze and communicate those findings, but the value comes from the insights you bring.
That’s the difference between covering a topic and contributing something to it. This is why your content strategy needs to start with your specific audience and your expertise.
Why experience and expertise matter more than ever
If you’re working on SEO or AEO, you’ve probably heard of E-E-A-T. It stands for Experience, Expertise, Authoritativeness, and Trustworthiness. It’s Google’s model that AI uses as signals to decide whose answers to trust.
Experience: Have you actually done the thing you’re writing about? Can you share first-hand observations, lessons learned, or examples from real work?
Expertise: Do you understand the subject deeply enough to explain it accurately, interpret the evidence, and offer useful guidance?
Authoritativeness: Is there a reason people should recognize you or your organization as a credible source on this topic?
Trustworthiness: Are your claims accurate? Do you cite sources? Are you transparent about your methods, limitations, and conclusions?
E-E-A-T is a useful framework for thinking about what makes your content credible and helpful. It’s especially relevant in an environment where AI can generate polished, confident-sounding content with very little effort. These qualities don’t automatically guarantee rankings or AI citations, but they help you build the kind of content people can evaluate, trust, and use.
Here are some best practices in this model:
- Include author bylines and bios with relevant credentials and real experience
- Cite primary, authoritative sources (research, standards, recognized industry orgs) and link to them
- Use specific data, examples, and case studies to support your claims (avoid generic statements)
- Include published and updated dates on posts
Make your content quotable, not just crawlable
This is one of the biggest opportunities for brands investing in AEO. Traditional SEO has always told us to make our content accessible to search engines, target relevant queries, and earn visibility in search results. Those fundamentals still matter. But in AI-powered search, there’s another question worth asking:
Have we given AI systems a useful reason to reference our brand?
Look at the difference between these two statements:
“Email marketing is an effective way for B2B companies to nurture leads and improve engagement.”
“We analyzed 14,000 B2B marketing emails and found that the messages generating the strongest engagement shared three common characteristics.”
The first statement is generic. It could come from any email marketing website. The second statement offers a specific finding and a reason to investigate further. It gives readers something concrete to learn from and potentially reference. That’s what I mean by making content quotable.
You want your content to include clear answers, meaningful insights, specific examples, credible evidence, and original information that contributes to a conversation.
You also want your website to make that information easy to understand. Use descriptive headings, answer questions directly, define important terms, cite your sources, and clearly identify the people and organizations responsible for the content.
The goal isn’t to write for an algorithm at the expense of your audience. It’s to create content that genuinely helps your audience and is clear enough for search engines and AI systems to understand.
What should you do differently?
You don’t need to stop using AI. You need to be more intentional about how you use it.
Here are five practical ways to put all of this into action.
1. Start with real audience questions.
Before you write a blog post, identify the question your audience is actually trying to answer. What are your salespeople hearing on discovery calls? What questions come up during onboarding? What misconceptions keep prospects from moving forward? What are customers struggling to understand?
Use those insights to determine what you should create.
AI can help organize the research, identify gaps, and develop an outline. But your understanding of the audience should guide the direction.
2. Build a point of view before you build the post.
Don’t ask AI to write an article about a topic and hope it comes back with a unique perspective. Decide what you want your audience to understand before you start drafting. What do you believe about this topic? What conventional advice do you disagree with? What have you learned from experience? What would you tell a customer that you wouldn’t find in a generic industry article?
Give AI that context and use it to help communicate your thinking.
Your point of view doesn’t have to be controversial. It does need to be clear, relevant, and supported by your experience or evidence.
3. Look for opportunities to create original evidence.
Original research is one of the clearest ways to differentiate your content, but it doesn’t have to mean commissioning an expensive industry-wide study. Start with what you already have access to.
- Analyze customer or campaign data you can legitimately use.
- Interview internal subject matter experts.
- Survey customers about their experiences.
- Identify patterns across projects or engagements.
- Publish case studies that explain the problem, approach, and results.
- Share lessons learned from experiments, including what didn’t work.
Be transparent. Don’t imply that a small or biased dataset represents an entire industry. The value isn’t in inventing a statistic. It’s in doing the work to uncover something meaningful.
4. Edit for insight, not just readability.
AI-generated content can sound polished without having anything useful or valuable to say. A grammar check won’t fix that, and neither will swapping out a few words to make the article sound more human. Look at your content and ask yourself:
- Does this teach the reader something useful?
- Is this supported by evidence?
- Are we repeating common advice without adding context?
- Have we included our own experience, analysis, or perspective?
- Could another company publish this exact article with almost no changes?
- What would make this piece more useful to our specific audience?
That last question is especially important. If your article could belong to any company in your industry, it probably needs more of your company’s thinking.
5. Measure value, not just volume.
Publishing more content does not equal a solid content strategy. Track the metrics that help you understand whether or not your content is contributing to business goals. Depending on the purpose of the content, that might include organic traffic, qualified conversions, engagement from target accounts, branded search demand, or visibility in relevant AI-generated answers.
The goal isn’t to win the publishing race. It’s to make your content investment count.
What does this mean for your SEO and AEO strategy?
The watermarking of AI-generated content doesn’t mean you need a separate content strategy for every LLM. It reinforces the importance of building a strong foundation.
That means making your website easy to crawl and understand. Create content that answers real questions. Demonstrate your expertise. Cite credible sources. Keep your information accurate and current. Build your reputation beyond your own website. And, most importantly, give your audience something they can’t get anywhere else.
For AEO, that means thinking beyond whether your page contains the right keywords. Your content should provide a clear, credible, and distinctive answer that would be useful in a conversation. You can’t guarantee that a particular LLM will cite your brand. You can, however, make your content more useful, your expertise more apparent, and your positioning easier to understand.
That’s a much better investment than trying to outsmart a watermark detector.
Let’s all promise to get back to providing our own thoughtful and valuable content vs. creating a bunch of copy-and-paste content from AI. Because when every company can generate a blog post in seconds, the advantage doesn’t belong to the company that publishes the most content. It belongs to the company that has something meaningful to say.
AI and the search landscape
Read More
